At present, domestic access control systems perform identity verification mainly by card-type devices, finger print and human face identification technology.
However, card-type devices can be easily lost and forgotten, and thus it is not only inconvenient for identity verification, but also imposes security risks. If a person with a bad intention obtains the card, he may implement theft. Current access control systems often adopt fingerprint identification and human face identification. Fingerprint identification can overcome the defects of the card-type products. However, around 5% people cannot be identified by fingerprints, and such identification methods by touching can spread infection easily. Furthermore, during season changes or dry seasons, fingerprint identification also has the problem of requiring several attempts so that the identification can be successfully done. The human face identification is limited by light, postures and accessories. When performing human face comparison, the facts that the user got shaved, is wearing eye-glasses or has an expression change may lead to the failure of the identification by the system and great inconvenience to the user identity verification.